AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
3,026 Commits 3 Branches 0 Tags
eb4053e3fecaeb5dd3952dfaafc93957d08038c2
Commit Graph

869 Commits

Author SHA1 Message Date
Aaron Enye Shi eb4053e3fe Guard the OCML rounded operations instead 
Instead of commenting all these functions out, guard the functions with a macro OCML_BASIC_ROUNDED_OPERATIONS.


[ROCm/hip commit: e59c33250a]
 2018-11-06 16:32:14 +00:00
Aaron Enye Shi 1c6f47ef55 Remove non-working non-default-rounded math apis 
In ROCm-Device-Libs, they have dropped the non-default-rounded versions of add, sub, mul, div, sqrt and fma. Therefore, ocml has removed the rte, rtp, rtn, and rtz counterparts. This will remove the same math APIs in HIP for _ru, _rd, _rn, and _rz.


[ROCm/hip commit: 2b108a2cfd]
 2018-11-05 22:34:16 +00:00
Siu Chi Chan 1159b4aa05 Move the global arrays for hip malloc/free 
from a header into a source file such that
there's only an unique copy in an executable
and prevent wasting static memory on the host

Change-Id: Id5b62766f77809c8d7b47892cb7149c490dcbdb9


[ROCm/hip commit: 0ff408a56c]
 2018-11-01 16:20:35 -04:00
Jeff Daily c28043bc4f typedef struct hipFuncAttributes now C compatible. 
Fixes #591.
Fixes #694.


[ROCm/hip commit: bb447dd76e]
 2018-10-31 14:12:45 +05:30
Maneesh Gupta ac071f5390 Merge pull request #714 from aaronenyeshi/remove-static-hip-device 
Remove libhip_device.a static library

[ROCm/hip commit: 0b86b421c1]
 2018-10-31 14:08:29 +05:30
Maneesh Gupta 3986e47f8d Merge pull request #713 from scchan/master-hcc_early_finalize_support 
Support more than one device code blobs in a single .kernel section

[ROCm/hip commit: 7f3e7b5334]
 2018-10-31 14:04:06 +05:30
Aaron Enye Shi 5bbd632888 Remove libhip_device.a static library 
Move remaining function definitions from device_util.cpp to hip_runtime.h header. We can now remove the static library completely as device_util.cpp was the last file part of target hip::hip_device .


[ROCm/hip commit: 817b27d530]
 2018-10-25 19:52:35 +00:00
Alex Voicu 147a6bab43 hipLaunchKernel, hipLaunchParm are deprecated, and shall be removed. 
[ROCm/hip commit: e4181b85be]
 2018-10-25 13:32:17 +01:00
Maneesh Gupta 5333a22ccc Merge pull request #723 from mangupta/fix_double_shfl 
Fix logic for double variants of __shfl*

[ROCm/hip commit: f6f160fa6b]
 2018-10-25 06:01:38 +05:30
Maneesh Gupta dd8ebc9a97 Fix logic for double variants of __shfl* 
Change-Id: I604f00b54cf4bd9c5f26ca6fa680fca5e9629417


[ROCm/hip commit: 19404e603d]
 2018-10-24 12:39:09 +05:30
Maneesh Gupta 90d15c13a1 Make HIP functional again with HCC from ROCm 1.9.x 
Change-Id: I214acdfd0b79dcf783993e44fe31baee64fd4dc3


[ROCm/hip commit: 0703a2d0f0]
 2018-10-24 10:41:56 +05:30
Maneesh Gupta a648f1c48e Merge pull request #705 from ROCm-Developer-Tools/feature_minimal_changes_for_hc_next 
Feature minimal changes for hc next

[ROCm/hip commit: 4a00b244a3]
 2018-10-19 06:58:31 +05:30
Siu Chi Chan 7ce37f3a08 Support more than one bundles in a single .kernel section 
When compiling with Early Finalization enabled in HCC,
the resulting .kernel section of the host object now may
contain more than one device code bundles.  This is to
teach the HIP runtime to correctly extract all the
bundles from the .kernel section.


[ROCm/hip commit: 30ce25e627]
 2018-10-18 17:00:27 -04:00
Alex Voicu b8b15ce792 Dumb workaround is still needed, so add it back. 
[ROCm/hip commit: 5ccaf2fa7d]
 2018-10-18 15:33:46 +01:00
Alex Voicu de15c6988d Re-sync with upstream. 
[ROCm/hip commit: fe959f7bd7]
 2018-10-18 12:27:03 +01:00
Maneesh Gupta 0dff12a615 Merge pull request #688 from aaronenyeshi/fix-sinf-cosf-ocml 
Use sinf and cosf from ocml device libs

[ROCm/hip commit: 1a5025c57e]
 2018-10-18 16:39:20 +05:30
Maneesh Gupta 1a8b4ce037 Merge pull request #692 from whchung/hip-reinit-take2 
HIP program state re-initialization logic (take 2)

[ROCm/hip commit: d133493669]
 2018-10-18 12:06:41 +05:30
Maneesh Gupta 63e6cd228b Merge pull request #703 from mangupta/stream_create_with_priority 
Implementation for stream priority

[ROCm/hip commit: c24b06fa0a]
 2018-10-17 10:53:43 +05:30
Maneesh Gupta a58a408e71 Merge pull request #702 from aaronenyeshi/fix-missing-irif-lib 
Replace IRIF fences with atomic_work_item_fence

[ROCm/hip commit: dbe4431d98]
 2018-10-17 10:53:27 +05:30
Maneesh Gupta 83e56b2ed4 Add missing hipHostRegister flags on nvcc path 
Change-Id: I69f09204d9c544935104d4168ab8d3626666a623


[ROCm/hip commit: 64d1cf86b7]
 2018-10-15 15:30:24 +05:30
Alex Voicu 69b0baf37f Minimal should mean minimal. 
[ROCm/hip commit: 5312336ce2]
 2018-10-11 00:21:41 +01:00
Alex Voicu d68ed45708 Address Aaron's comments 
[ROCm/hip commit: 3e4dbd32a1]
 2018-10-11 00:03:01 +01:00
Alex Voicu 0d30e15646 Merge branch 'master' of https://github.com/ROCm-Developer-Tools/HIP into feature_minimal_changes_for_hc_next 
[ROCm/hip commit: 4bc40551b5]
 2018-10-10 11:44:09 +01:00
Alex Voicu 6d0b50e9b9 Re-sync with upstream. 
[ROCm/hip commit: ca375cb8c5]
 2018-10-10 11:43:49 +01:00
Maneesh Gupta b35590db18 Implementation for stream priority 
- Requires ROCm 1.9.x or higher
- Requires HCC with PR#886 merged

Change-Id: Id7c95ea091ee610e80c9ad815f1cb989cba570ca


[ROCm/hip commit: da64156fb2]
 2018-10-05 16:27:46 +05:30
Aaron Enye Shi 188df279a9 Replace IRIF fences with atomic_work_item_fence 
[ROCm/hip commit: 0787f74ac2]
 2018-10-04 21:47:28 +00:00
Aaron Enye Shi ff1c7e9d27 Fix hip_vector_types.h for long long vectors 
There was a missing long in the declaration for [u]longlongN types.

[ROCm/hip commit: 5dd35576f6]
 2018-10-03 13:57:52 -04:00
Wen-Heng (Jack) Chung 4e04cf3d2a HIP program state re-initialization logic 
This commit is to support kernels dynamically loaded thru means such as
dlopen() after HIP runtime initializes.


[ROCm/hip commit: dab1a0f9db]
 2018-09-26 19:48:47 +00:00
Aaron Enye Shi 19d76382d6 Use sinf and cosf from ocml device libs 
Using llvm_amdgcn builtin fails to produce accurate values, we should move to using the ocml device library versions.


[ROCm/hip commit: 77c07d4118]
 2018-09-25 19:31:39 +00:00
Maneesh Gupta 70f9bd87c6 Merge pull request #614 from ROCm-Developer-Tools/fma 
Add overloading resolution functions for fma

[ROCm/hip commit: 3d67c9f952]
 2018-09-20 13:38:03 +05:30
Yaxun Sam Liu f6318119a9 Silent warnings about duplicate static keyword 
static is already in __DEVICE__, so should be removed.


[ROCm/hip commit: a5c961e26c]
 2018-09-19 10:39:45 -04:00
Yaxun Sam Liu 2d4040f75d Add fma function with float and _Float16 arguments 
[ROCm/hip commit: bd622a4b4a]
 2018-09-19 09:59:33 -04:00
Yaxun Sam Liu dcc99ba5c6 Fix build failure of hipTestHalf and hipTestIncludeMath for hip-clang 
[ROCm/hip commit: cf184460e9]
 2018-09-18 21:00:15 -04:00
Maneesh Gupta bec821768d Merge pull request #672 from iotamudelta/fp16_fix 
Only LLVM6 and higher contain the necessary intrinsics.

[ROCm/hip commit: 9ee70fca8a]
 2018-09-18 08:43:33 +05:30
Maneesh Gupta 4eba3cdf5c Merge pull request #674 from mangupta/fix_dtests_on_nvcc 
[dtests] Fix hipTestClock, hipTestNew, hipTestGlobalVariable, hipSimpleAtomicsTest & hipTestIncludeMath tests on nvcc path

[ROCm/hip commit: 32787fa1fc]
 2018-09-18 07:50:52 +05:30
Maneesh Gupta e4f978c9d6 Merge pull request #677 from yxsamliu/fix-launch-decay 
Fix hipLaunchKernelGGL for hip-clang

[ROCm/hip commit: 5cf281071d]
 2018-09-18 07:50:37 +05:30
Yaxun Sam Liu 3d60af4f35 Disable device code for gcc in hip_memory.h 
These device code should only be seen by HCC or hip-clang. They causd build failure
for HIP-VDI runtime and should be disabled for gcc.


[ROCm/hip commit: cdfd82f1de]
 2018-09-17 16:50:42 -04:00
Yaxun Sam Liu 694b6e105a Fix hipLaunchKernelGGL for hip-clang 
Do not decay function pointer type of the kernel argument passed to hipLaunchKernelGGL
and hipLaunchKernel, otherwise some type information is lost which may cause
type inference failure for the template.

This issue caused compilation error of FeatureLPPooling in Caffe2/PyTorch and this patch
fixes that.


[ROCm/hip commit: fc228c7ea6]
 2018-09-17 11:20:41 -04:00
Maneesh Gupta fc3c4aca00 Add mappings for __clock* in nvcc_detail/hip_runtime.h 
Change-Id: Ibcecf52f3e69298268d921efc036090544fa0ed0


[ROCm/hip commit: cef5261fa9]
 2018-09-17 15:23:30 +05:30
Alex Voicu 312d29ebaf Align with HC Next. 
[ROCm/hip commit: c6720e882b]
 2018-09-17 11:50:29 +03:00
Maneesh Gupta 4a671abf13 Merge branch 'master' into support-malloc 
[ROCm/hip commit: 66f863d1f3]
 2018-09-17 10:17:25 +05:30
Maneesh Gupta aef0245f90 Merge pull request #650 from ROCm-Developer-Tools/hip-clang-new 
Support placement new in hip-clang

[ROCm/hip commit: cb348421d7]
 2018-09-15 11:21:01 +05:30
Maneesh Gupta c8a776bfba Merge pull request #665 from aaronenyeshi/fix-min-funcs 
Use templates for min to prevent ambiguity

[ROCm/hip commit: 8fe4e22b19]
 2018-09-14 13:21:38 +05:30
Aaron Enye Shi fd9194fb16 Fix Tensorflow ambiguous min issue 
[ROCm/hip commit: 6b811ca6d1]
 2018-09-13 23:16:20 +00:00
Johannes M Dieterich b0602d501a Only LLVM6 and higher contain the necessary intrinsics. 
[ROCm/hip commit: cf12a9c049]
 2018-09-13 13:55:43 -05:00
Maneesh Gupta 5b2ca452c1 Merge pull request #669 from ROCm-Developer-Tools/feature_automatic_cast 
Remove potential for mismatch between runtime passed actuals and defined formals

[ROCm/hip commit: aed5ad31ba]
 2018-09-13 07:54:22 +05:30
Maneesh Gupta 2b59aea015 Merge pull request #661 from yxsamliu/add-empty-printf 
Add empty printf for hip-clang

[ROCm/hip commit: 411e53a665]
 2018-09-13 07:54:03 +05:30
Aaron Enye Shi 4d265c6df0 Avoid AMP-retrict call to CPU-restrict 
[ROCm/hip commit: 894cbdd749]
 2018-09-12 14:54:31 +00:00
Alex Voicu f6100dae08 Remove potential for mismatch between runtime passed actuals and defined formals. 
[ROCm/hip commit: cdfea3ef7b]
 2018-09-12 10:30:48 +01:00
Maneesh Gupta c1d8dc86ff Merge pull request #664 from lcskrishna/master 
added __host__ to float2half and half2float functions.

[ROCm/hip commit: 8249cf037b]
 2018-09-12 14:50:01 +05:30